Crawlers - Segmented multilegged rovers

dicktice · 2002-11-29 16:35:09

(Updated 14 December 2002.)

Crawling Rovers: Has anyone of you not seen the multilegged insect-like like robots being shown on TV? They just seem to "flow" over obstacles. I would think them ideal, in principle, for unmanned Mars surface surveys--scaled-up to cope with the size of rocks strewn about in the terrain we have been shown so far. Being squat, they couldn't be tipped over, or toppled into gullies or from cliffs...if programmed to hesitate with their trailing-segment legs gripping surfaces already traversed by their leading-segment legs--until images of what had been encountered was transmitted (via whatever intermediate mother-vehicle to Earth) and interpretated by the remote-presence operator.
Watching these things move on TV--their top "scuttling" speeds are really prodigious--has pretty well convinced me that wheels will never match the match the advantages of "centipedal" propulsion, for Earth-based remote presence rover missions.
The current generic term adopted for these things is "Polypedal Robots." (Any suggestions for something more catchy?) They are being adapted from actual insects at the University of California at Berkely's PolyPEDAL Laboratory, by Professor Bob Full and a team of graduate students.
I invisage centipede-like rovers with segmented bodies--four or six legs per segment--with three or more per rover; individually powered by batteries allied with solar cells; remotely attachable and detachable for (a) maintenance replacement or (b) changing mid-mission demands requiring add-on manipulators, drills, tote-buckets and /or sensors; and interconnected wirelessly via digital FM radio (to avoid connectors) such that control can shifted end-for-end or to any segment between.
I hope this suggestion may stimulate a discussion about the pros and cons of unmanned crawler-type rovers. (For more background, search: "Polypedal Robots.")

19 replies since June 24 2002,11:09 < Next Oldest | Next Newest >

Pages: (2) < 1 [2] >

soph · 2002-12-12 20:39:01

. Being squat, they wouldn't fall over. Or topple into gullies or from cliffs (say) if programmed to hesitate with their trailing feet gripping surfaces already traversed by their leading feet...until images of what was ahead had been transmitted (via whatever) to Earth for evaluation.

its funny, i had a lego robot set that could do just that. well, almost. it had a photocell that made the robot stop when it came to a black line. simple progamming, for kids. well, my guess is that we could make a device extend, say, a yard in front of its legs, attached to the body, that sent sonar beems to the ground. when it came to a point where the ground dropped significantly, it would stop.

CalTech2010 · 2002-12-14 18:23:00

How would the robot differentiate between a 1000 ft cliff dropping off into nothingness, and a gently sloping crater wall? I mean, if you wanted to maneuver right up to the edge you could find out, but with sonar you would only be able to tell where the edge is, not its relative position. ???

soph · 2002-12-14 21:00:31

thats not true. sonar is able to locate position and depth. thats how subs work.

if you stick the radar a yard in front, you can stop at the edge.

CalTech2010 · 2002-12-16 19:55:36

I get what you're saying now. I was thinking you would use a Sonar line-of-sight to see through the edge and what was on the other side.

Even with a sonar out front by a meter, how would you navigate down "steps"? A washed out hillside with ledges in the path would look like a drop still, even though it is easy to get through. You couldn't keep asking mission control to summarize the gradient, because there's a 12 minute overlapse in transmission.

I think we should have accurate maps of the surface, and some kind of GPS system that is highly accurate, and a real-time video monitor operating in near-light frequencies.

soph · 2002-12-16 22:21:01

if youre operating it from a base, you dont have to wait 12 minutes. it is also fairly easy to have it compute the gradient itself (if its a yard in front, it can be programmed to slow down, shift its sonar arm to see whats further in front, and decide whether to proceed or go...this is all pretty simple programming), so it can be semi-independent.

CalTech2010 · 2002-12-17 15:57:08

Are you suggesting we use crawlers as human-assistance machines once we get to Mars, or use them as Earth-launched explorers? I was thinking you intended the latter.

If you wanted it to be human-assistance, what would you use it for? It would be too slow and tedious to use for mapping, and tight, dangerous places would required millions of calculations. If you wanted to use it as cargo, you would need to account for weight and balance, which increases the complexity of going down a slope.

soph · 2002-12-17 16:58:30

no...robots used for human assistance dont need to be so slow. cargo robots, when they are needed, later on, can be built on site to be as large as needed. robots now are only so small and slow because a large rover is hard to control, and the robots used now are operated from a long distance, constantly checking back for new instructions.

on mars, a steady video link could be established, and the rover could constantly be given new programming, chips, etc, based on what they find. they can be adapted. replacements could be built to adapt. the rovers now are built slowly for caution-if it goes, its gone. a little metal on site could turn into a new rover. hell, a little plastic could too. these rovers would be ideal for mapping.

and they could be used either way, there is no reason why they couldnt.

Josh Cryer · 2002-12-18 16:10:11

Yeah, vehicles on Mars would be much faster than automonous robots, just think operating a big bulldozer.

I don't like this crawler design, though. Legs are inefficient compared to wheels. Also, keeping the whole vehicle stable, without everything bouncing up and down would prove difficult. With wheels, we just have shock absorbers. With legs we would actually have to calculate the exact position of the main leg joint with regards to the rest of the legs. Wheels are much simpler than legs, repairing them would be quite trivial. Legs on the other hand would have to at least have two joints (a wheel only having one berring), for the whole thing to be reasonably stable.

I think legs are a novel idea, but probably won't be so great on Mars.

CalTech2010 · 2002-12-18 16:12:44

Yes, but the scale of maping would be so small. I suppose if you needed VERY accurate maps of the surface, it would be okay. If you were in charge of surverying the surface of mars, would you use robots or engineers and surveyors? I suppose the two could be used in unison.

soph · 2002-12-18 16:27:55

they could be used in unison. besides, radar mapping could be used to deliver a miles wide radius. do you think we've stepped over every square inch of the world? no. but we've got nearly every pebble on earth mapped. the robot doesnt need to be directly over a point to map it.

robots could be given all kinds of cool equipment to do things people couldnt on site without large amounts of equipment.

CalTech2010 · 2002-12-18 21:04:27

do you think we've stepped over every square inch of the world? no. but we've got nearly every pebble on earth mapped.

That's because we mapped it from space, not using multi-million dollar machines to do the job a $65,000 engineer can do. I'm not saying they wouldn't have their uses (I think they would be great at assisting human surveys), however, I think mapping isn't one of them.

soph · 2002-12-18 21:11:40

we had the world mapped before we were in space. and its not all by people. you cant have somebody walking around mapping. it just wont work. a robot, like i said, can map for a several mile radius. their computer can be programmed to calculate gradients, draw a 3d image, and anything else youd want. robots would be far more useful for mapping than people. a robot could see for kilometers underground for water, metals, minerals, life signs, geothermal vents, etc.

they could perform tasks automatically, like picking up sampling rocks. this is all at no risk to the people, who are needed for other things. i dont think you understand just how valuable robots are. take a high end computer programming course, or engineering course, and youll see.

CalTech2010 · 2002-12-18 21:14:42

I don't doubt the usefulness of robots, I just wouldn't trust it for major surveying. I could see its use in a low-grade, approximation mapping mission, but not for any kind of construction surveying.

soph · 2002-12-18 21:21:41

you could have it do any grade of detail you want. to have it do low grade work would be a waste, and pointless. thats pretty much backwards. robots could do highly detailed, large area mapping. why wouldnt you want them to?

if they found something promising, and you got paranoid as to their reliability, send someone in to check.

CalTech2010 · 2002-12-18 21:24:21

Would the robots send information back to base to be analyzed and catalogued, or would it compute it right then and there and send back the finished product?

I was thinking that robots couldn't map in great detail with a several kilometer range, but that's how we mapped the ocean floor, right?

soph · 2002-12-18 21:30:53

could do it either way, depending on what you wanted. it could flag data that matched certain criteria, and send it back "high priority."

seti has a thing called seti-at-home, where people download a program, and it uses their PC to compute "data chunks" of nearby space. it uses this flagging system, and the scientists pay special attention to flagged data. flagging data is also used in error reporting. its fairly common.

and yes, it could do highly detailed analysis...thats probably what they used for the ocean floor.

CalTech2010 · 2002-12-19 23:05:05

I suppose you couldn't use crawlers for construction anyway. The cargo size involved would be too big to efficiently operate under that leg size.

soph · 2002-12-20 12:09:12

i dont think thats true. you could brace the vehicle as much as you want. you could also strengthen the legs with high tension cables, which can hold up huge amounts of weight. legs provide an advantage, in that they could extend and contract, so a robot with legs could go as high or low as it wants, and in many directions.

i would think that construction would require a variety of autonomous robots, but overall, it would be better than using people, because it would be safer, cheaper, and it would free people up for other jobs.

CalTech2010 · 2002-12-22 15:54:14

Let's not put the cart in front of the horse... have you seen the "latest" crawler on a scale of that size at MIT(?)? It had six legs, and it takes about 90 seconds to move one leg out in front of the others and stabilize it. I would much rather use a rover for the non-rocky areas of Mars

soph · 2002-12-22 16:03:55

You could always use a combination. treads would work too. by the time we get to mars, the leg system will probably be much more advanced.

CalTech2010 · 2002-12-23 00:26:22

What? What would cause NASA to focus research on legs between now and Mars when they crap themselves just playing with the ISS!

soph · 2002-12-23 07:16:18

Playing with? You dont have any clue about the ISS. Just because people say its not the best expenditure, it isnt useless. People have to stop using the ISS as the universal whipping boy.

dicktice · 2002-12-23 11:06:08

Let's not put the cart in front of the horse... have you seen the "latest" crawler on a scale of that size at MIT(?)? It had six legs, and it takes about 90 seconds to move one leg out in front of the others and stabilize it. I would much rather use a rover for the non-rocky areas of Mars

Not so: the latest centipede-like robots just seem to "flow" over objects, including verticals, with the advancing part supported by the trailing parts still on the horizontal. Progressive leg action is automatically (not cognitively) coordinated, forward or backward, and they can be made to just creep along or go like hell. This is the way to go! (See University of California at Berkeley, Robot Labs.)

CalTech2010 · 2002-12-23 13:26:20

I've seen the little insect crawlers, and I have to admit they're pretty cool for a $250,000 set of LEGOs. Soph and I were talking about using a crawler for construction purposes, which would be a slow cumbersome beast.

And I wouldn't trust a Berkley student as far as I could throw one...

New Mars Forums

Announcement

#1 2002-11-29 16:35:09

Re: Crawlers - Segmented multilegged rovers

#2 2002-12-12 20:39:01

Re: Crawlers - Segmented multilegged rovers

#3 2002-12-14 18:23:00

Re: Crawlers - Segmented multilegged rovers

#4 2002-12-14 21:00:31

Re: Crawlers - Segmented multilegged rovers

#5 2002-12-16 19:55:36

Re: Crawlers - Segmented multilegged rovers

#6 2002-12-16 22:21:01

Re: Crawlers - Segmented multilegged rovers

#7 2002-12-17 15:57:08

Re: Crawlers - Segmented multilegged rovers

#8 2002-12-17 16:58:30

Re: Crawlers - Segmented multilegged rovers

#9 2002-12-18 16:10:11

Re: Crawlers - Segmented multilegged rovers

#10 2002-12-18 16:12:44

Re: Crawlers - Segmented multilegged rovers

#11 2002-12-18 16:27:55

Re: Crawlers - Segmented multilegged rovers

#12 2002-12-18 21:04:27

Re: Crawlers - Segmented multilegged rovers

#13 2002-12-18 21:11:40

Re: Crawlers - Segmented multilegged rovers

#14 2002-12-18 21:14:42

Re: Crawlers - Segmented multilegged rovers

#15 2002-12-18 21:21:41

Re: Crawlers - Segmented multilegged rovers

#16 2002-12-18 21:24:21

Re: Crawlers - Segmented multilegged rovers

#17 2002-12-18 21:30:53

Re: Crawlers - Segmented multilegged rovers

#18 2002-12-19 23:05:05

Re: Crawlers - Segmented multilegged rovers

#19 2002-12-20 12:09:12

Re: Crawlers - Segmented multilegged rovers

#20 2002-12-22 15:54:14

Re: Crawlers - Segmented multilegged rovers

#21 2002-12-22 16:03:55

Re: Crawlers - Segmented multilegged rovers

#22 2002-12-23 00:26:22

Re: Crawlers - Segmented multilegged rovers

#23 2002-12-23 07:16:18

Re: Crawlers - Segmented multilegged rovers

#24 2002-12-23 11:06:08

Re: Crawlers - Segmented multilegged rovers

#25 2002-12-23 13:26:20

Re: Crawlers - Segmented multilegged rovers

Board footer